Output: 20W / 8 Ohms
Frequency Range: (TBD)
Low Noise, Solid State Power Supply
There is a long standing vogue for SETs. The A Number One problem with these amplifiers is that they don't produce enough power to drive any but the most efficient speakers to adequate volumes. Types commonly used for this purpose: 45, 2A3, the TV vertical deflection triodes, or trioded 6V6s -- despite the reputation for sonic excellence -- typically don't produce more than a couple of watts. The only exception is the 300B, an audio power triode developed the same year as the 807, one of the first beam pentodes. This wrote the end of further audio triode development. Given the right speakers, these amps sound lovely, to be sure. If you don't have the special, high efficiency speakers, you are pretty much limited to using such audio triode amps with headphones. The Wolverine, however, has sufficient output power to drive most speakers to comfortable listening volumes.
The only new power triodes to appear subsequent to the development of the beam pentodes were designed as series pass regulators, TV vertical deflection amps, or RF finals. Most of these series pass regulators are duals. (What possibly compelled them to put two triodes in the same bottle, knowing that the end users would be paralleling them anyway escapes me.) Being that these weren't designed for audio, less attention was paid to balance between sections. This can pose problems for sonic performance. The vertical deflection finals are all in the same power class as the 2A3, since vertical deflection duty isn't as demanding as horizontal deflection duty. These can make fine audio finals since the requirements are the same: low-u factors, low rp's to damp the vertical deflection coils, and excellent linearity for good, linear vertical scanning. The vast majority of the RF triodes are of the "zero bias" variety with high u-factors for high power sensitivity and ease of drive. As with high-u, small signal types, the rp's run high. They also tend to produce more higher order harmonics. This limits the SET designer to the vintage triodes, or the TV vertical deflection types, such as the 6CK4.
There is, however, one power triode that is an exception to the rule: the 845 -- a DHT with a decent PD rating. The 845 was developed in the 1930s to serve as an audio final for AM modulators. As with any audio power triode, the 845 has the desirable low u-factor and low rp. This type was designed for the convenience of being able to share the HV power supply between the RF finals and modulator. Even though the 845 was intended for AM modulators, where fidelity isn't always desirable, necessary, or legal, it does offer excellent sonics.
The Wolverine offers interstage transformer coupling, and no global NFB. Under the right conditions, transformer coupling can offer excellent, low distortion performance.
The design of the gain stages is straight forward. The first preamp is the small signal half of the 6DR7. This type was designed to serve as the vertical deflection oscillator and amplifier for B & W TVs. The small signal section is similar to half of a 6SL7, and features excellent linearity with a decent gain margin. Since high voltage is not a problem with this project, a passive plate load large enough for good linearity with adequate current sourcing is not a problem. The second section is a low-u triode power amp with a low rp. This makes it a particularly good candidate for driving an interstage transformer. The low rp allows for good bass response, since the primary inductive reactance becomes a larger percentage of plate resistance. This stage also uses fixed bias in order to avoid the distortion problems that can arise from capacitor bypassed cathode resistors. Capacitive currents in the cathode circuit are capacitive currents in the plate circuit which includes the transformer. This can lead to some nasty distortion. The plate current of the large signal section is stabilized with a DC servo. It operates with sufficient plate current in order to drive the 845 control grid for good high frequency and overdrive behaviour.
The final is a single ended, Class A1 design. Getting the VPP up keeps the current demand low. Since the OPT must carry the static plate current, the design of a good single ended OPT becomes a good deal more formidable than the usual push pull OPT without the DC core magnetization problem. Core magnetization at lower plate voltages and higher plate currents seems to come with sonic penalties when using 845s. The best THD estimates occur at the higher operating voltages. The final uses fixed bias for improved linearity. The actual design could not be simpler.
The Wolverine can output some 20W of audio power: more than enough to drive almost all speakers to comfortable listening levels. The inclusion of a stiffer driver stage also allows for a couple of extra watts, and improved clipping behaviour during the occasional high level transients that occur in all music programs.
Copyright © 2009 All rights reserved